The year of 2015 was marked by the historical event of the first direct detection of
gravitational waves (GW) by the Laser Interferometer Gravitational-Wave Observatory
(LIGO). This Nobel Prize winning accomplishment was only possible after the original
facility (first-generation LIGO observatory) suffer major updates enhancing its sensi-
tivity by one to two orders of magnitude. Even with the unprecedented capabilities of
the updated LIGO (formally a second-generation, hereby 2G), we were only possible to
detect GW from events of mergers of binary systems containing black holes (BH) and/or
neutrons stars (NS), as these coalescences are among the brightest astrophysical sources
of GW. Given the success of 2G detectors, there are on-going plans for the construction
of even more sensitive detectors, the so called 3G detector. The increase in sensitivity
will allow not only to expand our detection horizon, but will also enable the detection
of the fainter GW sources.

    This Thesis is dedicated to the study of the morphology and underlying physics
behind prospective targets of 3G detector. In particular we are interested in two types
of source: the first one refers to the very exciting (albeit very speculative) possibility
of BHs lacking an event horizon. This violation of such a fundamental prediction of
General Relativity (GR) laws leads to a train of repeating GW pulses in latter stages
of the signal emitted by a BH merger event. Dubbed as GW echoes, these pulses are
expected (if they do exist) to be at least ≈ 1 order of magnitude smaller than the GR-
expected signal. In this Thesis it is shown a detailed analysis about the morphology of
this signal leading to the conclusion that the presence of echoes can be very-well tested
with 3G detectors and how this endeavor will constitute the most stringent test of GR’s
validity in the proximity of the event horizon. The formal description of echoes in this
project relies on the regime of validity of the linear perturbation theory of BHs.

    The second kind of GW source of central interest to this document is of a less hypothetical nature. 
Comparatively weaker than the BH/NS coalescences by 1 to 2 orders of magnitude, 
GW emitted by gravitational-collapse driven events are understood to be only detectable in 2G 
detectors if they take place within the Milky Way. In this work we pay close attention to the 
GW radiation of accretion-induced collapse (AIC) of White Dwarfs (WD). AIC are collapsing 
events that, contrarily to Supernova Ia, do not undergo over a thermonuclear explosion. 
During the execution of this work, we performed Numerical Relativity (NR) simulations to predict 
the GW strain emitted by our AIC models and used them to access their expected signal-to-noise 
ratio (SNR) in future 3G detector.

	This Thesis has as final intent to help in painting the full picture of the sciences
cases for 3G ground-based GW detectors. To do so, we studied two families very distinct
families of GW sources, and by doing so we provide arguments in favor of using these
planned detectors as means of detecting not only fainter sources but also access potential
deviations from GR. We strongly believe that these two projects (in collection to many
others in the literature) show how much more new knowledge will be made available by this novel
detector network.